Ramsar COP7 DOC. 20.1
"People and Wetlands: The Vital Link"
7th Meeting of the Conference of the Contracting Parties
to the Convention on Wetlands (Ramsar, Iran, 1971),
San José, Costa Rica, 10-18 May 1999
Ramsar COP7 DOC. 20.1
Technical Session V:
The framework for regional and international cooperation regarding wetlands
Shared Wetlands and River Basins of the World:
Preliminary Findings of a GIS Analysis
Report prepared in support of Ramsar COP7 Technical Session V
by Dr Brian Groombridge, World Conservation Monitoring Centre
Ramsar sites and international borders
Ramsar sites and international catchment basins
Ramsar sites and basin vulnerability
Wetlands areas and international catchment basins
Recommendations for further study
Annex 1: Catchment basin boundaries
Annex 2 : Key to country codes
1. The purpose of this project was to highlight opportunities for international cooperation over management of wetland sites. The study was based on two assumptions:
i. that cross-border, near-border or coastal sites may be at risk from transboundary effects, and
ii. that sites within shared catchment basins may be similarly at risk, particularly where most of the basin and territory of more than one country are located upstream of the site.
Geographic Information System (GIS) technology was used to analyse sets of spatial data with respect to these assumptions.
2. Of a total 955 Ramsar sites included in this study, 92 (10%) are located in part or entirely within 10km of an international border, and thus in principle may be subject to transboundary influences, such as water abstraction or drainage. See Table 1.
3. There are nine cases where Ramsar sites in adjacent countries are in contact along the international border (or appear to be at the scale of this analysis). These adjacent sites (19 in all) provide the opportunity to be managed cooperatively as integrated transboundary Ramsar sites, and this should be given high priority if it is not occurring already. See Table 2.
4. A small number of Ramsar sites, 35 in all, are both within 10km of a border and located on the coast of their respective countries and so potentially at increased risk from transboundary influences. See Table 3.
5. Almost one third of the Ramsar sites considered in this study, 267 (28%) in total, are located within catchment basins shared between two or more countries (152 of 227 basins in this analysis). The need for international cooperation is in principle likely to rise with the number of countries sharing any given basin, and with the area and number of countries upstream of that site. See Tables 4 and 5.
6. Of these 267 sites within shared basins, 191 are within basins ranked as significantly vulnerable on the basis of previous analysis (criteria: low naturalness and high water stress; WCMC, 1998). Basin vulnerability assessment can provide a tool for prioritising management intervention. See Table 6.
7. Extensive areas of wetland habitat are present within most shared catchment basins. See Tables 7 and 8. Africa has a large number of basins shared between five or more countries, most holding very extensive wetland habitats, which would potentially benefit from international cooperation. Further analysis is required to make practical use of information on wetland habitats in relation to borders and basins.
8. A number of recommendations for improving the depth and precision of this analysis, and further increasing the potential conservation benefits, are outlined in this report.
9. This report outlines results of a project carried out for the Secretariat of the Convention on Wetlands (Ramsar, Iran, 1971) by the World Conservation Monitoring Centre (WCMC).
10. The purpose of the report is to demonstrate, on the basis of readily available information, situations where countries may have a responsibility under the Convention to cooperate in the management of wetlands and associated drainage basins.
11. The project is based on Geographic Information System (GIS) analysis of relevant global spatial datasets, some of which have been enhanced for the purpose. Results have been verified by direct inspection of paper plots of the data and of source tables. Data layers comprise: 1:1,000,000 world outline, with national boundaries; major world catchment basin boundaries; world river systems and lakes; Ramsar sites; wetlands (as plotted and classified for the Wetlands in Danger atlas, Dugan, 1993); world mangrove systems.
12. Results of this analysis should be treated as indicative, not definitive. The project is intended to demonstrate the actual or potential need for international cooperation, based on the generalised risk factors outlined below. No attempt has been made to assess the biodiversity value of the different wetland sites and areas, nor to collate field-scale information from the literature or gather new field data.
13. Most emphasis is on areas declared under the terms of the Ramsar Convention. This is because Ramsar sites are clearly delimited and identifiable areas, and the spatial data are well-suited to GIS analysis, but preliminary use has been made of spatial data on wetland habitat areas.
RAMSAR SITES AND INTERNATIONAL BORDERS
14. The purpose of this phase of the investigation was to determine which Ramsar sites are in contact with, or in close proximity to, an international border. The limit of the close proximity zone was arbitrarily set at a distance of 10 km; the analysis could be repeated using some other value (e.g., five or 50 km). Existing Ramsar sites were sorted into two groups: 1) those in contact with or close to an international border; 2) those not in close proximity.
15. The main working assumption in this case is that, other factors being equal, close proximity to an international border is likely to increase the susceptibility of the site to transboundary influences (e.g., drainage, groundwater pollution) and hence increase the potential need for international cooperation. A second assumption is that sites that are in contact across international borders must be highest priority for international cooperation, where this is not currently in place.
Procedures and data quality
16. Ramsar sites are represented by site boundaries where these are digitised, or by circles based on centroid coordinates and proportional to site area, where site boundaries are not digitised. WCMC has developed digital coverage of site boundaries for 692 sites (66 for the purpose of this project); only location coordinates are available for the remaining 263 sites. The basic list of sites, a total of 955, was retrieved from the Ramsar Convention website (list dated 27 October 1998). All Ramsar sites within 10km of an international border have been identified by using a GIS to relate the boundaries and proportional circles in this dataset to national boundaries.
17. Accuracy is considered satisfactory for the purposes of the project, given the global scope of the analysis. The fact that proportional circles are used to represent Ramsar sites lacking digital boundary data will cause some error in the distance calculations. This will probably be insignificant in most cases, but is liable to introduce significant error where sites are highly elongated.
18. Of the total 955 Ramsar sites considered here, 92 (10%) are situated in part or entirely within 10km of an international border (Table 1).
19. Nine wetland regions include Ramsar sites that are, or appear to be on the basis of data used in this analysis, in direct contact across international borders. These sites present an opportunity to be managed as integrated transboundary Ramsar sites, and international cooperation to this end should be treated as high priority (where not already in place).
20. These high priority sites are listed in Table 2. Some of these sites, e.g., Danube delta area, are also downstream sites susceptible to influence originating in other countries in upper parts of the basin in which they occur, and are also coastal sites and thus susceptible to cross-border marine impacts.
21. A significant number of near-border sites are also directly on the coast of their respective countries, and so are potentially at increased risk from transboundary factors. These sites are listed in Table 3.
Next steps needed
22. A set of Ramsar sites that may be at risk from transboundary factors, and therefore deserving of international cooperation, has been identified. Further analysis is required, using different and where possible more fine-scaled data to qualify potential levels of risk. For example, sites where human settlements around the international border are known to draw water from a shared groundwater resource, or are downstream of known sources of waterborne pollutants, may be regarded as at high risk. Information on the biodiversity value of different sites, and their importance to human communities, would also contribute to prioritisation of sites. Management regimes in adjacent Ramsar sites apparently in contact along an international border should be assessed with a view to improved harmonisation where necessary.
RAMSAR SITES AND INTERNATIONAL CATCHMENT BASINS
23. The purpose of this phase of the investigation was to determine which Ramsar sites are known to be located within international catchment basins, i.e., basins shared between two or more countries. Existing Ramsar sites were sorted into three groups:
i. those within shared basins,
ii. those within basins situated within a single country, and
iii. those located outside catchment boundary dataset used in this analysis (see below) and which therefore could not be attributed to shared or non-shared basins.
24. The first working assumption is that in principle, and ignoring the complexity of case-specific factors, the need for international cooperation over any given Ramsar site increases in parallel with the number of countries sharing the basin in which it is located.
25. A second assumption is that the need for cooperation will rise according to the number and area of countries within the basin upstream of any given site. In other words, sites at the mouth of the drainage system, where they may be subject to the influence of activities taking place in all of the upstream countries, are likely to be higher priority for international cooperation than upstream sites close to the watershed of the basin.
Procedures and data quality
26. Both the datasets used in the previous exercise (Ramsar sites, national boundaries) were analysed against the major world catchment boundaries dataset. This procedure allows all Ramsar sites situated within shared basins to be identified, and any given basin to be ranked according to the number of countries sharing it.
27. The shared basins part of the analysis is limited by the coverage and quality of the catchment basin boundary dataset. Global level datasets are typically based heavily on computed flow patterns over landscape represented by a digital elevation model (DEM); this is sufficient for many purposes but does not necessarily capture actual flow patterns, especially in regions with even topography and low slope. Sources differ over the exact boundaries of most catchment basins, and in some cases therefore over the number and identity of countries sharing them.
28. We have assembled a basin boundary dataset that combines elements from different sources and which has been matched so far as possible with conditions on the ground as represented in conventional small- to medium-scale printed maps (see Annex for note on sources). The dataset includes 227 principal basins, including the worlds larger systems and a geographic sample of smaller catchments.
29. This dataset has been expanded for the purpose of the present project, starting with the set containing 151 basins used for a catchment condition analysis in WCMC (1989). However, many transboundary basins are very small indeed at world scale (see Anon., 1978), and within the limits of the present project it has not been possible to include these small basins. A total of 214 major international basins were recognised in 1978 (Anon.), and since that date, fragmentation of previous country units (e.g., former USSR) has increased the number of shared basins. The dataset used here includes 152 international basins (13 of which are shared by five or more countries). Although this suggests that a significant number of shared basins are likely to be missing from the present dataset, the great majority of these are likely to involve two countries only. The present analysis does include all the larger basins, most of those involving three countries, and probably all those involving more than three countries.
30. Of the 955 Ramsar sites considered in this study, 267 (28%) are located within the international drainage basins included in our dataset (Table 4). Of these 267 sites, 62 are also within 10km of an international border.
31. Table 5 shows a number of Ramsar sites that appear to be relatively high priority for international cooperation, because they are located at or near the mouth of a large international river system, or have a number of other countries located upstream of them. This indicative list of downstream sites, intended to highlight the issue, has been compiled by direct inspection of maps without using rigorous quantitative criteria. Two examples are illustrated in Maps 1 and 2.
Next steps needed
32. A set of Ramsar sites that may be at risk from transboundary factors, in particular from those factors liable to impact the drainage network within international catchment basins, has been identified. In principle, these sites are deserving of international cooperation. Further analysis is required, using different and where possible more fine-scaled data to qualify potential levels of risk. For example, a site in one country that is downstream of known sources of waterborne pollutants may be regarded as at high risk. Information on the biodiversity value of different sites, and their importance to human communities would contribute to the further prioritisation of sites.
RAMSAR SITES AND BASIN VULNERABILITY
33. The objective of this part of the study was to investigate the feasibility of using additional data and assessment methods to identify possible priorities among the set of sites judged to be at possible risk; i.e., sites that also appear on independent evidence to be at risk might be regarded as priority for further investigation or management intervention.
Procedures and data quality
34. WCMC (1998) carried out a trial global assessment of the naturalness of catchment basins, and the level of stress they are likely to come under as a result of increasing water demand. It can be argued that Ramsar sites within international basins ranked as vulnerable may themselves be at significant risk.
35. The vulnerability of basins was estimated by a combined measure of present naturalness and future pressure on water resources. Naturalness was estimated by a GIS analysis (by Rob Lesslie, Australia); this entailed measuring the distance of all grid points from mapped indicators of human impact and access (roads and other transport, buildings and settlements, etc.). These grid values were averaged over each basin, with the assumption that dense infrastructure and access is an indicator of low naturalness. This has been shown to be a good surrogate for other aspects of human influence on landscapes. Water resource vulnerability is based on an analysis by Raskin et al. (1997) in which countries were scored according to estimated future pressures on water resources.
36. Table 6 lists the 191 Ramsar sites that occur within basins assessed as higher vulnerability in WCMC (1998). As indicated in the table, 68 of these sites are also within 10km of an international border. Map 3 shows the position of a number of sites at the mouth of the Danube, a high priority complex of sites by all the criteria used in this overview study.
Next steps needed
37. This preliminary assessment of basin vulnerability could be much refined and would then provide a clear rationale for determining priorities for action. For example, countries sharing the Niger system are at high risk of increasing water stress; wetlands within the system are vulnerable and, on this basis, high priority for international cooperation. In contrast, although the Amazon system as a whole is shared between a large number of countries, it is at very low risk of water stress, and the basin as a whole is ranked at low vulnerability in WCMC (1998). However, parts of the system are severely impacted by waterborne pollutants, and drainage-specific factors of this kind need to be included in future more detailed vulnerability assessment.
WETLAND AREAS AND INTERNATIONAL CATCHMENT BASINS
38. The intention of this exercise was to assess the need for international cooperation directed at broad areas of wetland habitat, rather than discrete wetland sites as exemplified by areas declared under the Ramsar Convention.
Procedures and data quality
39. An attempt was made to use the same procedures for areas of wetland habitat as were applied to Ramsar sites, outlined above. However, the results are in general of less value because the available spatial data relating to wetland habitats are in their present form not suitable for global analysis.
40. The wetland dataset was collated primarily in support of a major wetlands atlas (Dugan, 1993). It was based heavily on regional wetland directories produced over several years by IUCN The World Conservation Union, WWF, UNEP and Wetlands International, supplemented by a number of national map sources. The data and the resulting maps were described as "the most comprehensive and accurate assessment of the worlds wetlands compiled to date" (Dugan, 1993), and it remains the case that there is no other global dataset more suitable for the present analysis.
41. The main limitations of this dataset for the present application are: 1) most of the wetland areas are either very large in extent, or highly fragmented, or both, and 2) in most cases, a simple classification of wetland type (based on Ramsar Convention categories) is the only attribute linked with the spatial data, and even this classification is not entirely consistent between continents. In North America the two main classes are for areas with 25-50% wetland and 50-100% wetland. More attribute data are available for most wetland areas in Africa, where a site name is also stored. Thus, in this project, it has been possible to determine from GIS analysis the relative area of wetland types within each international basin (or country), but without further analysis this information has limited application.
42. Spatial data on mangrove occurrence were derived from a global dataset collated at WCMC, as used in a recent atlas of mangroves (Spalding et al. 1997). A layer representing peatland soil, as a potential indicator of bog and mire habitats, was used, derived from the FAO Digital Soil Map of the World (Version 3.0).
43. Although analysis of the proximity of Ramsar sites and international borders (outlined above) appears useful, this approach did not yield results open to ready interpretation in the case of wetland habitats. Table 7 shows the area of each main wetland type within each of the international catchment basins in this analysis. Larger basins tend to hold greater areas of wetland. Basins shared by five or more countries make up most of the drainage in continental Africa; these basins also hold very extensive areas of wetland habitat. The great extent of swamp forest in the Amazon system, and of fresh water marsh in the Ob and Parana systems, are noteworthy. The latter wetland type is widely correlated with peatlands, as in the Ob and Amur basins. Significant areas of mangrove are present around the mouth of many shared catchment basins, particularly extensive in the Ganges-Brahmaputra, Kapuas and Orinoco systems. Table 7 summarises areas of peatland and mangrove in international basins.
Next steps needed
44. Although the very limited use of the wetland area dataset in the present study did not yield results of immediate use, there is clear potential to extend this phase. In particular, a GIS could be used to analyse the extent to which wetland areas are covered by designated protected areas, and gaps could readily be identified, in terms of wetland type, country and catchment basin. This could be carried out using datasets already available at WCMC. Gap analysis of this kind would generate results with immediate application.
45. Further analysis would probably require improvements to be made to the wetland area dataset, in part by adding attribute data and in part by refining the typology used.
46. Although some use was made of data on mangrove occurrence, it was not possible within the limits of the present study to consider coastal marine wetland habitats in significant detail. Spatial data on coral reef areas are available. These habitats could be treated in more detail in a future study.
RECOMMENDATIONS FOR FURTHER STUDY
47. A number of important steps that would collectively further improve the scope, focus and conservation benefits of this preliminary GIS analysis have been noted above. These are summarised in the paragraphs below.
48. A set of Ramsar sites has been identified that may be vulnerable to transboundary factors by virtue of their location in relation to international borders or within international catchment basins. Further analysis is required, using different and where possible more fine-scaled data, to qualify levels of risk to these sites. Data on the biodiversity value of different sites; on their importance to human communities; on occurrence of shared groundwaters, or on sources of waterborne pollutants, are among the categories of information that would contribute to this further analysis.
49. Management regimes in adjacent Ramsar sites apparently in contact along an international border should be assessed with a view to improved harmonisation where necessary.
50. Sites identified as at potential risk from transboundary effects that are also located within catchment basins identified as vulnerable on the basis of other evidence would clearly be candidates for priority attention. A preliminary high level assessment of basin vulnerability (WCMC, 1998) has been applied in this study and appears to provide a useful tool for suggesting priorities. This preliminary assessment should be developed further by including more detailed system-specific information.
51. An additional GIS study should be undertaken to analyse the extent to which wetland areas are covered by designated protected areas. Gaps in coverage could readily be identified, in terms of wetland type, country and catchment basin, and such results would have immediate applications.
52. Beyond this gap analysis, further use of the wetland areas dataset would probably require improvements to be made to the data, in part by adding attribute data and in part by refining the typology used.
53. Spatial data on coastal marine wetland habitats should be treated in more detail in a future study.
Anon. (Centre for Natural Resources, Energy and Transport of the Department of Economic and Social Affairs, United Nations). 1978. Register of International Rivers. Water Supply and Management. 2: 1-58. (Special Issue). Pergamon Press.
Dugan , P. (ed). 1993. Wetlands in Danger. Mitchell Beazley, in association with IUCN The World Conservation Union. Reed International Books Ltd., London.
Raskin, P., Gleick, P., Kirshen, P., Pontius, G. and Strzepek, K. 1997. Water Futures: Assessment of Long-range Patterns and Problems. Background Report #3 of Comprehensive Assessment of the Freshwater Resources of the World. Stockholm Environment Institute.
Spalding, M., Blasco, F. and Field, C. D. (Eds.). 1997. World Mangrove Atlas. The International Society for Mangrove Ecosystems, Okinawa, Japan.
World Conservation Monitoring Centre. 1998. Freshwater Biodiversity: a preliminary global assessment. Groombridge, B. and Jenkins, M. WCMC - World Conservation Press. Cambridge, UK.
ANNEX 1: CATCHMENT BASIN BOUNDARIES
The catchment basin boundaries used were based initially on the global dataset on the GlobalARC CD-ROM made available by CERL (the US Army Corps of Engineers Construction Engineering Research Laboratories). This is generated from a relatively coarse elevation model. Data for North America, Africa and Europe were replaced with improved boundaries generated by the United States Geological Survey (USGS) and made available at: http://edcwww.cr.usgs.gov/landdaac/gtopo30/hydro/index.html . The most inclusive basin boundaries, i.e., the entire drainage system passing through one river mouth (or delta region) to the sea, were used in the analysis, selected to provide a reasonable sample from each continent. A very few internally draining systems were also included. In addition, the major catchment boundaries were inspected by eye against appropriate paper maps, and adjustments made.